[gdal-dev] Re: Proposal: gdal_muladd.py

[Antonio Valentino]

Hi Jukka,

Il giorno Tue, 2 Feb 2010 07:20:52 +0000 (UTC)
Jukka Rahkonen <jukka.rahkonen at mmmtike.fi> ha scritto:

> Greg Coats <gregcoats <at> mac.com> writes:
> 
> > 
> > I would prefer to obtain in one step the absolute value of the
> > difference
> between two pixels, in part because
> > I do not see the meaning of a negative pixel value. In my example,
> > I do not
> see meaning to a signed byte value
> > for a pixel channel being red=-51. Greg
> 
> Hi,
> 
> If somebody is doing for example change detection by comparing
> satellite images from different dates, it is not at all the same if
> the result is 99-150=-51  or 150-99=51.
> 
> -Jukka Rahkonen-

I completely agree with you.

Anyway, IMHO the point here is not to decide if we have to
*modify* the diff function so that it returns abs(b1-b2).
Greg asked to add a new one ( abs(b1-b2) ) to the *base set* of pixel
functions.
Stated that there is no technical difficulty to do this I'm wondering
what is the Frank's mind about the question.

In my opinion the criterion should be to have only a minimal set of
pixel functions in the library and write a support tool for all other
needs.

Another criterion could be to have a wider set of pixel functions in
the library and to include all the ones that are considered *of
general utility*.
Greg's request makes perfectly sense, of course, with respect to this
criterion.

I think I will add to the patch abs(b1-b2).
It will be GDAL developers, if they ever will consider to integrate the
patch, to take a decision about which pixel function to include and
which not.


Best regards


-- 
Antonio Valentino